The NCI/NINDS Brain Tumor Progress Review Group (PRG) has highlighted that the available rodent model systems for brain tumors do not accurately represent the biology of human brain tumors. As a consequence the PRG has designated the development of new animal models as a high priority for research. We propose to investigate the molecular cytogenetics of spontaneous canine brain tumors, as a first step towards supporting the role of the domestic dog as a suitable animal model that will offer valuable insights into the genomics of brain neoplasia. Domestic dogs are organized into more than 350 phenotypically distinct genetic isolates ('breeds') characterized by unique constellations of morphology, behavior, and susceptibility to specific diseases, including cancers of the brain. Humans and dogs have similar physiology, share extensive genome homology with a high degree of preserved gene order, and are exposed to the same environment. Combined with the fact that the incidence and lifetime risk of naturally occurring brain tumors in dogs differ markedly between dog breeds, this offers a unique opportunity to identify molecular factors that define risk and prognosis. In previous studies, we have shown that evolutionarily conserved chromosome aberrations exist in human and canine hematopoetic cancers. Significantly, we also identified chromosomal aberrations in dogs for which corresponding changes have not been reported in humans. We hypothesize that these regions of the canine genome may contain cancer-associated genes that have thus far been intractable in studies of human families and tumors due to the genetic 'noise' present in human populations. Our long-term goals are to identify genetic lesions that account for the origin of brain cancers and that influence disease progression and prognosis. For this project, we propose to utilize the emerging canine genome assembly to select a panel of 3,000 BAG clones and generate a genome-wide microarray with 1Mb resolution. We will then conduct array-based comparative genomic hybridization (array-CGH) analysis of 50 canine brain tumors as a means by which to identify recurrent regions of copy number change. Comparative bioinformatics will be used to identify genes located within these regions of genomic imbalance that warrant further investigation. Characterization of these genetic loci will help to improve the definition of molecular risk and prognosis of human and canine brain tumors, with a potentially major impact on human health.